← Blog
By Engine Simulator Team20 min read

Why Do Diesel Engines Make So Much Torque? We Took the Turbo Off to Find Out

Long stroke, high compression, or the turbo? We measured a Cummins 6BT against two gasoline V8s, then removed each lever to see which one torque follows.

An inline-six diesel long block on a red engine stand with its head removed and six bores showing, a large cast-iron turbocharger sitting detached on a wooden workbench in the foreground, an aluminium small-block V8 on a second stand behind it, cold window light from the left and a warm work lamp on the right

A Cummins 6BT displaces 5.9 litres and, in our simulator, makes 607 N·m at 1,714 rpm. A Ford Modular 4.6 V8 displaces 4.6 litres and makes 395 at 3,918. Ask a truck forum why diesel engines make so much torque and you will have five answers inside a minute: the long stroke, the compression ratio, the turbo, the fuel, the slow burn. Every listicle on the first page of results names the same five and weighs none of them.

We can weigh them. The 6BT in our roster is calibrated to the real engine's 1998 rating within 1.3%, and its turbo, its stroke and its compression ratio are each one field in a config file. So we took them off one at a time and watched where the torque went.

Key Takeaways

  • Per litre, the turbodiesel leads by a margin, not a multiple. In our simulator the 6BT makes 12.97 bar of BMEP at its torque peak against the LS7's 11.61 and the 4.6's 10.78: +12% and +20%.
  • Take the turbo off and the diesel falls to 9.24 bar and 73.5 N·m per litre, below both gasoline V8s. The turbo is 29% of its torque, and that is the air side alone.
  • The diesel's 120 mm stroke, given to the 4.6 at equal displacement, created no torque: −3.8% at the peak, and the peak moved 379 rpm down the tach.
  • The diesel's 17.5:1 compression, given to the 4.6 with knock switched off, is worth +29.5%, about 3.2% per point compounded. Pump gasoline cannot collect it.
  • The rev ceiling is a burn fixed in time, not the stroke. Scaled with rpm, the burn costs 13% of the torque at 3,000 and 21% at 4,000, while mean piston speed at redline is 12.0 m/s against the LS7's 23.7.

On this page: Quick answer · Pressure times displacement · Lever one: the turbo · Lever two: the stroke · Lever three: compression · Where the torque lives · Why it cannot rev · Why the truck feels stronger · What we do not model · Try it in the simulator · FAQ · How these numbers were made

Quick Answer: Why Do Diesel Engines Make So Much Torque?

Because they are large, turbocharged, and make their torque at 1,700 rpm rather than 4,500. In our simulator a 5.9 L Cummins 6BT makes 600 N·m at 1,500 rpm, where a 7.0 L LS7 makes 590 and a 4.6 L Ford makes 370; per litre the diesel leads the two gasoline V8s by 12% and 20%, and the whole of that lead comes from a turbocharger working on a 17.5:1 engine. Take the turbo away and, per litre, the diesel trails both.

In these measurements the fuel's energy and the long stroke contribute nothing; the stroke moves torque down the rev range without adding any.

Torque Is Pressure Times Displacement

Every argument about diesel torque is an argument about one equation. Brake mean effective pressure is the constant pressure that, acting on the piston through one power stroke, would deliver the measured torque; for a four-stroke it is

BMEP = 4π · T / Vd

with torque in N·m and displacement in m³ (Wikipedia, Mean effective pressure, retrieved 2026-09-10; Heywood, Internal Combustion Engine Fundamentals, ch. 2). Read it backwards and torque is BMEP times displacement times a constant. An engine has two ways to make more torque: be bigger, or make more pressure per litre. "Diesels make more torque" is therefore a claim about BMEP, and BMEP is a number.

Below is the number for the three engines this article measures, in our simulator. Each row is a 29-point wide-open sweep from idle to redline of the calibrated stock configuration. The Otto-cycle side of what the pressure is doing is in the Otto cycle explained.

BuildLitresCRBoost target (bar)Peak hpPeak N·mN·m per litreBMEP at peak (bar)N·m @ 1,500N·m @ 2,500N·m @ 3,000
Cummins 6BT (turbo_diesel_i6_5L9)5.88317.51.25212.1 @ 3,000607.2 @ 1,714103.212.97600.3527.0503.4
LS7 (v8_7L0)7.01111—501.5 @ 5,875647.6 @ 4,52592.411.61590.4612.6624.4
Modular 4.6 2V (sohc_v8_4L6)4.6019.4—269.0 @ 5,054394.6 @ 3,91885.810.78369.8382.3388.3

The folklore, stated as a number, is 12.97 against 11.61. Per litre the turbodiesel is ahead of the LS7 by 12% and of the 4.6 by 20%, which is a real lead and not a doubling. The LS7 out-torques it outright, 648 N·m to 607, by being 1.1 litres bigger. That is the equation doing what it says.

The real 6BT is a 102 × 120 mm turbocharged inline-six fitted to Dodge Ram pickups through the 1990s (Wikipedia, Cummins B Series engine, retrieved 2026-09-10). Our preset is calibrated to the 1998 high-output rating of 215 hp at 2,700 and 597 N·m at 1,600, and the committed calibration reads −1.3% on power and +0.2% on torque against it; the LS7 and the 4.6 are calibrated to their published ratings the same way. An engineering reference on BMEP sets out the same yardstick across engine types (EPI Inc., BMEP: a performance yardstick, retrieved 2026-09-10); it is the right lens for the question, because it strips displacement out. So far the diesel is doing what everyone said. The question is which lever is doing it.

Lever One: The Turbo, Removed

The same 6BT, same 17.5:1, same 120 mm stroke, same injectors, with the turbocharger keys deleted from the config and nothing else touched:

BuildBoost target (bar)Peak hpPeak N·mN·m per litreBMEP at peak (bar)N·m @ 1,500N·m @ 2,500N·m @ 3,000
6BT stock1.25212.1 @ 3,000607.2 @ 1,714103.212.97600.3527.0503.4
6BT, turbo removed—133.7 @ 3,000432.4 @ 1,47373.59.24431.8345.2317.3

The turbo is worth 175 N·m at the peak, 29% of the stock figure, and 78 hp. Without it the diesel's BMEP is 9.24 bar and its torque per litre is 73.5: below the LS7's 92.4, below the 4.6's 85.8, below both gasoline engines in the roster it was meant to be out-muscling. A naturally aspirated 5.9 L diesel in this model is a 432 N·m engine, and a 5.9 L version of either gasoline V8, on its own per-litre figure, would beat it.

That is the contrarian case Curbside Classic makes in its title, that gasoline engines intrinsically make more torque and power than diesels (Curbside Classic, Why gasoline engines intrinsically make more torque and power than diesel engines, retrieved 2026-09-10). Litre for litre and with the turbo removed, our row supports it; the reputation, on these numbers, is the turbocharger's. The mechanism is in how a turbocharger works: a diesel runs unthrottled, with excess air, and meters torque by fuel, so every kilogram of air the compressor adds is a kilogram the injector can burn against.

One honesty note on that 29%. A diesel in this model has no per-injector ceiling; its fuel scales with the air available, so the number above is the air side of the turbo's contribution, complete and exact. A real 6BT is also capped by its injection pump and injectors, which is why our calibration notes record the boost target as low on purpose: with nothing to cap the fuel, the published rating has to be matched by limiting air instead. The real engine's turbo is worth whatever its pump lets it be worth. Ours is worth 29%.

Lever Two: The Stroke, at Equal Displacement

"Long stroke makes torque" is the most-repeated line in the folklore (TractorByNet, Diesel torque difference mostly myth?, retrieved 2026-09-10), and it comes with an intuition: a longer stroke is a longer crank throw, a longer lever arm, more torque from the same push. The equation above has already told us what is wrong with it. At equal displacement a longer stroke means a smaller bore, and the force on the piston is pressure times bore area. The lever arm grew by the same ratio the piston shrank. Torque is BMEP times displacement, and neither changed.

Measured, on the 4.6. We rebuilt it with the diesel's 120 mm stroke and the bore that keeps it at 4.601 L, 78.1 mm (the bore and stroke calculator does the arithmetic), held the rod ratio, kept the same heads, the same compression and the same volumetric-efficiency curve, and ran both with knock switched off (octane: 200) so the comparison is thermodynamic and not a knock model's opinion:

BuildBore × stroke (mm)Peak hpPeak N·mN·m per litreBMEP at peak (bar)N·m @ 1,500N·m @ 2,500N·m @ 3,000
4.6, stock geometry90.2 × 90.0269.0 @ 5,054394.6 @ 3,91885.810.78369.8382.3388.3
4.6, 120 mm stroke, same litres78.1 × 120251.1 @ 5,054379.4 @ 3,53982.510.36366.2374.5377.9

Peak torque −3.8%, at 3,539 rpm instead of 3,918. Torque at 1,500 rpm −1.0%. Peak power −6.6%. The long stroke moved the torque peak 379 rpm down the tach and created no torque anywhere on the curve; what it did create was mean piston speed, which at 6,000 rpm rose from 18.0 to 24.0 m/s, and friction rises with piston speed (Heywood, ch. 13), which is how our friction law charges it, so the top end paid for it.

The stroke relocates torque without manufacturing any. If the diesel's low peak is what people mean by "diesel torque", the stroke is part of that, and this test shows the direction. It is not the magnitude.

The caveat this row must carry: the VE curve is a preset input and we held it. A real long-stroke rebuild has a smaller bore, so smaller valves and a lower, earlier VE peak, and a real one would lose more at the top than this. Nothing in that direction helps the long-stroke case.

Lever Three: Compression, With Knock Switched Off

A diesel compresses air, not mixture, and lights it by injecting fuel into air already hot enough to ignite it, which is why it runs compression ratios a gasoline engine cannot (Wikipedia, Diesel engine, retrieved 2026-09-10). The Otto-cycle ideal says thermal efficiency rises with compression ratio, so more of the same fuel becomes pressure on the piston. How much? We gave the 4.6 the diesel's 17.5:1, again with knock off, because on pump fuel a 17.5:1 gasoline engine reads through the knock derate and the question here is the thermodynamics:

BuildCRPeak hpPeak N·mN·m per litreBMEP at peak (bar)N·m @ 1,500N·m @ 2,500N·m @ 3,000
4.6, stock, octane 2009.4269.0 @ 5,054394.6 @ 3,91885.810.78369.8382.3388.3
4.6 at 17.5:1, octane 20017.5351.2 @ 5,243510.9 @ 4,107111.013.95472.6491.4500.1

Peak torque +29.5%, peak power +30.6%, over 8.1 points of compression: about 3.2% per point, compounded. At 17.5:1 the naturally aspirated 4.6 reads 13.95 bar, more than the turbocharged diesel's 12.97. This is a gasoline engine borrowing the diesel's compression without paying for it, and on pump fuel it cannot borrow it: the end gas detonates first. What engine knock is explains why a premixed charge has that limit and a compression-ignition charge does not, and the compression-ratio sweep, published today, measures where the gain ends when knock is switched back on.

So the compression ratio is real and it is large. It is also the reason the turbo lever looked as modest as it did: the 6BT's 17.5:1 is why its BMEP is competitive at all with only 1.25 bar of boost target. The two levers multiply, and the compression alone, with the turbo removed, does not get the diesel above the gasoline engines.

Where the Torque Lives

Go back to the 1,500 rpm column of the first table. At 1,500 rpm the 5.9 L diesel makes 600 N·m; the 7.0 L LS7 makes 590; the 4.6 makes 370. The diesel's torque peak is at 1,714 rpm, the LS7's at 4,525, the 4.6's at 3,918. At 1,500 the diesel is within 1% of its peak and the LS7 is 9% below its own, and the diesel is winning outright with 1.1 fewer litres.

By 3,000 rpm the order has reversed: the LS7 is at 624 and still climbing, the diesel is at 503 and has been falling since 1,714. That fall is the shape our calibration reproduces from the published curve, the fitted VE curve and turbo sizing, and the ceiling at 3,000 is the factory redline typed into the preset. What the diesel has, then, is displacement, a turbo on a 17.5:1 engine, and a torque peak at 1,700 rpm rather than 4,500. The peak's location is the part no listicle measures, and it is most of what the driver feels.

Why a Diesel Cannot Rev

The usual answer is the stroke again: long stroke, high piston speed, the parts cannot take it. The numbers say otherwise. At its 3,000 rpm redline the 6BT's mean piston speed is 12.0 m/s. The LS7 reaches 23.7 m/s at 7,000 and the 4.6 reaches 18.0 at 6,000. The diesel is stopping at half the piston speed the LS7 runs at every day. The stroke is not what ends its rev range.

In a teaching model, time ends it. Diesel combustion is mixing-controlled: fuel burns as fast as the spray can find air, and that rate is set in milliseconds by the injector and the turbulence in the chamber, not in crank degrees (Heywood, ch. 10; Wikipedia, Diesel engine, retrieved 2026-09-10). A burn that takes a fixed number of milliseconds takes twice as many crank degrees at twice the rpm. A gasoline flame front speeds up with the turbulence rpm brings, so its burn in crank degrees stays put (Heywood, ch. 9).

Our model's burn is a Wiebe curve in crank degrees, 62° on the 6BT, fitted to the published curve near 1,700 rpm; its cost does not grow with rpm, which is a gasoline-like assumption. To see what a time-limited burn does, we scaled that 62° with rpm so that it always took the 6.08 ms it took at 1,700, lifted the redline to 5,000 for the run only, and held one point per speed, once with each burn:

rpmBurn, fixed (°)Burn, time-scaled (°)N·m, fixedN·m, time-scaledΔhp, fixedhp, time-scaled
1,5006254.7591.9605.2+2.2%124.7127.5
2,0006272.9627.8607.4−3.2%176.3170.6
2,5006291.2610.7559.3−8.4%214.4196.4
3,00062109.4571.3495.6−13.3%240.7208.8
3,50062127.6534.9439.2−17.9%262.9215.9
4,00062145.9535.8422.1−21.2%300.9237.1

At 3,000 rpm a burn that takes the same 6.1 ms it took at 1,700 costs the diesel 13% of its torque; at 4,000, 21%. Below 1,700 the same rule gives it 2% back, because the burn is shorter in degrees there, which is the low-rpm side of the same coin. Power with the time-scaled burn is 237 hp at 4,000 against 209 at 3,000: a thousand more rpm buys 28 hp. That is the shape of a diesel's ceiling. Revving it harder returns less and less, until the expansion stroke is mostly over before the fuel is.

Both caveats below matter. First, above 3,000 rpm the 6BT preset is outside its calibration; the fixed-burn column past 3,000 is the model's behaviour, not the engine's, and we do not report it as a finding. The comparison is between two runs of the same model at the same rpm, and that is what makes it usable. Second, the real ceiling is also the injection pump's delivery curve and the governor, neither of which is modelled, and the reciprocating mass, 1.35 kg per cylinder against the LS7's 0.85, which our friction law charges only through piston speed. A real 12-valve stops where its P7100 pump says so. Ours stops where the preset says so.

Why the Truck Still Feels Stronger

You do not drive a dyno sheet. You have a foot, a tachometer and a gearbox, and the gearbox multiplies whatever torque the engine makes at the rpm it happens to be turning. At a 1,500 rpm cruise the diesel is making 600 N·m, within 1% of its peak. The LS7 is making 590, and to get to its 648 it needs to be at 4,525, which means a downshift, or two, and a wait. The 4.6 is making 370 with a 3,918 rpm peak it will reach later.

At the pedal, "diesel torque" means the torque is there already, at the rpm a loaded truck is turning, without asking. The per-litre lead is modest, and the bigger gasoline engine has more torque outright. A gasoline engine that peaks at 4,500 spends most of a towing day well below its best, and a converter's stall multiplication is what covers the gap (how a torque converter works measures that multiplication on this simulator). A diesel is at its best at 1,700 and the gearing is chosen to keep it there.

Add the turbo, which in this model is 29% of that 600, and the picture is complete. The truck feels stronger because it is large, boosted, and geared around a peak at 1,700 rpm. Every other explanation in the folklore is either a small effect, a relocation, or the turbo wearing a different hat.

What This Simulator Does Not Model

Every figure above is this simulator's, and it is a teaching model with known debts. Reading the numbers as facts about a real 6BT would be wrong in these specific ways:

  • No injection-pump curve and no governor in the sweep. The 6BT's power peak lands at 3,000 rather than the rated 2,700 for this reason; the real engine's power curve ends where the P7100 says so, and our dyno sweep deliberately ignores the governor. Our calibration notes record the discrepancy rather than tuning it away.
  • No per-injector smoke limit. A diesel here has no injector ceiling, so fuel scales one for one with air. The turbo-removed figure is the air side of the turbo's contribution only. A real 6BT is pump-limited as well.
  • Diesel fuel consumption is inverted in this model. The consumption ordering between diesel and gasoline is a known debt, so this article makes no claim about diesel efficiency, mpg or BSFC, and the energy-per-litre question in the FAQ is answered from published figures, not from the simulator.
  • Exhaust temperature is unmodelled. It follows the throttle lever rather than combustion, so nothing here is said about EGT.
  • The burn is a Wiebe curve in degrees, not a spray model. The time-scaled table emulates a mixing-limited burn by scaling one number; it is the right mechanism at the level of a teaching model, and the only claim it supports is the comparison between the two columns.
  • Above 3,000 rpm the 6BT is outside calibration, and the ceiling is an input. The model does not reproduce why the real engine stops at 3,000; it is told to. The fall-off from 1,714 to 3,000 is the fitted VE curve and turbo sizing that reproduce the published curve.

The full scope, and the rest of what we get wrong, is in how we build the simulator.

Try It in the Simulator

Three engines, one tachometer, the whole article by the seat of your trousers:

  1. Run the Cummins 6BT in the Truck class. Open the dyno and confirm the peak near 1,700; then drive it and watch how little of the tach you use. Cruise at 1,500 and floor it. The redline slider goes to 9,000, but past 3,000 this preset is outside its calibration, so read anything up there as the model's, not the engine's.
  2. Run the LS7 in the same class. Same cruise, same 1,500 rpm, same floored pedal, and notice where the tach has to go before anything much happens.
  3. Run the Modular 4.6, then open its spec editor and type in the diesel's stroke (120 mm) with the bore that holds 4.6 L, or its compression ratio (17.5). The second one lights the knock readout on pump fuel; that reading is the reason lever three belongs to the diesel.

Every curve here is produced by the physics, not looked up, so an edit you make is measured the way ours were. To build a diesel of your own geometry, use the build wizard.

Run the 6BT · Run the LS7 · Browse the full roster

Frequently Asked Questions

Why do diesels have low horsepower but high torque?

Because power is torque times rpm, and a diesel has nowhere to take the rpm. Our 6BT makes 607 N·m at 1,714 and peaks at 212 hp at its 3,000 redline; the LS7 makes 648 N·m at 4,525 and peaks at 501 hp at 5,875, from a torque figure only 7% higher. The horsepower gap is the rev gap. What limits the revs is the burn fixed in time, the pump and the governor, not the stroke.

Is diesel torque because of the long stroke?

Not in our measurements. Giving the 4.6 the diesel's 120 mm stroke at the same displacement lost 3.8% at the torque peak and 1.0% at 1,500 rpm, and moved the peak 379 rpm lower. At equal displacement a longer stroke means a smaller bore, so the lever arm gains what the piston area loses. The stroke moves torque down the rev range; it does not add any.

Does the turbo make the torque?

Most of the diesel's advantage, yes. In our simulator the 6BT's turbo is worth 175 N·m, 29% of its peak torque, and without it the diesel's torque per litre falls to 73.5, below both the LS7's 92.4 and the 4.6's 85.8. That is the air side only; a real 6BT is pump-limited as well, so the true share is whatever its injection pump permits. The compression ratio is the other lever, and it works by letting the turbo's air do more.

Why can't a diesel rev to 7,000?

Because diesel combustion is mixing-limited and takes a near-fixed time, so at higher rpm it occupies more of the expansion stroke. Emulating that in our model, a burn held at the 6.08 ms it takes at 1,700 rpm costs 13% of the torque at 3,000 and 21% at 4,000, and a thousand rpm from 3,000 to 4,000 buys only 28 hp. On a real engine the injection pump's delivery curve, the governor and the heavier reciprocating parts, 1.35 kg per cylinder on our 6BT against 0.85 on the LS7, add limits this model does not carry. Piston speed is not the reason: the diesel stops at 12.0 m/s, half the LS7's 23.7.

Does diesel fuel have more energy than gasoline?

Per litre, yes: diesel is denser and has a higher volumetric energy density than gasoline (Wikipedia, Diesel fuel, retrieved 2026-09-10). That affects fuel economy, not torque. Torque follows the air trapped in the cylinder and the pressure the burn makes on the piston, which is why a diesel with the turbo removed trails the gasoline engines despite the fuel. Nothing in this article measures consumption or efficiency, because our simulator's diesel consumption model is a known debt and we do not report it.

How These Numbers Were Made

It matters which kind of claim each figure is:

  • Physical principles that hold on real hardware: torque as BMEP times displacement, the lever-arm cancellation at equal displacement, compression ignition permitting ratios a premixed charge cannot, mixing-controlled combustion taking time rather than degrees. Corroborated by the cited sources, not derived from our code.
  • Simulator calibration: every torque, power, BMEP and per-litre figure in the tables. They describe three engines as this model calibrates them, the 6BT to the Cummins 1998 HO rating (−1.3% on power, +0.2% on torque, from our calibration notes) and the LS7 and 4.6 to their published ratings. Piston speed follows from the geometry; reciprocating mass (1.35 kg and 0.85 kg per cylinder) is a preset input, not a measurement.
  • Method: 29-point wide-open sweeps from idle to redline (full fuel rate on the diesel), each point settled before sampling, in fixed reference air of 1.000 bar and 293 K with no correction applied. Torque at 1,500, 2,500 and 3,000 is linear interpolation between sweep points. Every rebuild is the preset's input with one field changed and re-derived the way the preset builder derives it, so the injector, intake and turbo follow the change. The gasoline rebuilds run at octane: 200, which switches the knock model off; that is the mandated control, because a 17.5:1 gasoline engine on pump fuel would read through the knock derate and the question was the thermodynamics. The time-scaled burn is one held point per rpm with the burn duration set to 62 · rpm / 1700 degrees and the redline lifted to 5,000 for the run only. The knock model's random draw is seeded, so every figure repeats.
  • The rounded percentages (12%, 20%, 29%, −3.8%, +29.5%, 3.2% per point, 13%, 21%) are computed from the table figures shown, not from unrounded intermediates.
  • Reference peaks on the roster page come from a separate run on its own sweep grid, so a peak may land a few N·m and a few hundred rpm from the figures here. Same physics, different sample points; neither is more real.
  • No real-engine dyno figure appears above except the Cummins rating used as the calibration anchor. Where we describe what a real engine does, it carries a citation.

Every run reproduces from shipped presets: load one, open the dyno, and you should land on the same curve. If you do not, send it to us. The dyno correction calculator explains how to compare our reference-day figures against a real dyno sheet.

About this article

Written by the Engine Simulator Team, who build and calibrate the physics engine behind these numbers. We publish the model's limits beside its results, because a simulated figure is worth nothing without them. Found an error? Contact the team.

Sources and Further Reading

Related Reading